Guide to the Last Mile Table of Contents

WWW.MEGGER.COM

The word “Megger” is a registered trademark “Megger” is a registered The word

A Benchmark for VDSL Testing Guide to the Last Mile Table of Contents

Introduction...... 2 Resistive fault locator (RFL)...... 14 What is VDSL?...... 2 Load coil detection...... 15 Balance...... 16 Why the Old Ways Won’t Voice-band noise...... 16 Work with VDSL...... 3 Voice-band power influence...... 17 Basic Testing Procedure...... 4 Longitudinal balance...... 18 Characteristics of a Perfect Pair Stress test...... 18 for Installing VDSL2 Service...... 6 Noise...... 19 Factors That Can Cause a Pair to Voice-band spectrum analyzer...... 20 be Ineffective for VDSL Service ...... 7 Wide-band spectrum analyzer...... 21 Breakdown of Impulse noise ...... 22 insulation resistance...... 7 Open meter...... 25 Grounding ...... 8 Time Domain Reflectometer Bonding...... 8 (TDR)...... 27 Bridged tap...... 9 Bridged tap...... 28 Load coils...... 10 Steps to Turn up a VDSL Circuit.....30 Excess length...... 10 Troubleshooting VDSL...... 32 Impedance changes...... 10 Slow service/reduced data Unbalanced pair...... 11 throughput...... 32 Power company...... 11 Factors in Selecting a Tests to Effectively VDSL Test Set ...... 35 Deliver VDSL Service...... 12 AC volts...... 12 Megger HT1000 Copper Wire Analyzer...... 37 DC volts...... 12 Resistance...... 12

Author: Ed Rousselot, Megger Technical Sales Ed has been employed in the telephone industry since 1971 when he went to work for Southwestern Bell. He worked for telephone companies until 1984 when he took a position with a communications test equipment manufacturer. He has been in the industry since then in technical sales and support positions.

Ed holds a BS degree from Missouri State University and a MS degree from Oklahoma State University.

www.megger.com Guide to the Last Mile 1 Introduction How does Triple Play (television, voice telephone and internet access) work Telephone companies (telcos) have been generating increasing levels of on VDSL2? The International Standards Organization’s (ISO) open system revenue out of their twisted-pair copper outside plant (OSP) cables by interconnection (OSI) 7-layer protocol stack defines a hierarchy of duties offering high-speed data services. These services are, in most cases, delivered or tasks for establishing and maintaining a communications link across a from the telco’s central office over fiber to a remote terminal and, from this network so that the applications that reside at the top layer (email, IPTV, remote terminal, to the customer over what is called the “last mile,” a/k/a VoIP, etc.) will work. We do not have to worry about what these are called twisted-pair copper cable. The latest of these services is VDSL2. The stringent and the differences between layers and the different methods of achieving demands that VDSL2 place on this copper cable require specific tests to the task of a particular layer. What we need to know is that each layer “qualify” it as being able to carry VDSL2. Test equipment has been designed communicates with the layers directly above and below it and, through this, to qualify this last mile of twisted-pair copper for VDSL and to fix the VDSL with its corresponding layer at other points in the network to establish a service if something goes wrong with it. communications path.

This booklet is written as a guide to assist the telecom technician in Why the Old Ways Won’t Work with VDSL qualifying and maintaining twisted pair cable for VDSL service. This section will cover why the old ways of testing and installing services are Factors that a technician may want to consider when selecting a piece of test not conducive to having satisfied VDSL subscribers. equipment for qualifying twisted pair for VDSL service are addressed in the Telcos have deployed services on twisted-pair for years that required them Appendix. to make alterations to their POTS (plain old telephone service) lines. In the 1970s they had to remove load coils to accommodate analog subscriber What is VDSL? carrier. In the 1980s they had to remove bridged taps and load coils to offer VDSL is the acronym for Very high-speed Digital Subscriber Line. It provides a service called ISDN (Integrated Services Digital Network). Frame Relay faster data transmission over twisted pair, which allows for the provision debuted and the telco installer became involved not only in making sure of high-bandwidth applications such as Internet Protocol Television (IPTV), that the twisted-pair (the physical layer) was very clean, but he or she might Voice over Internet Protocol (VoIP), and very-high-speed internet access. have to “go up the protocol stack” to install and troubleshoot the service. VDSL works over twisted pair. Finally ADSL came along which further tightened the rules about what kind The first version, VDSL1, with its 1.2 MHz bandwidth and asynchronous data of twisted pair was required. Despite all these changes in technology, telcos rates of up to a theoretical 16 Mbps upstream and 52 Mbps downstream, could select from available vacant plant a twisted pair that would work for did not allow telcos to offer all the features that cable television companies these applications. offer, such as the ability to simultaneously view several HDTV programs. VDSL service is much touchier about what kind of twisted pair it will work VDSL2, on the other hand, utilizes bandwidth up to 30 MHz to provide over than any of the above services. Most ADSL1 service had the “blazing” theoretical data rates up to 100 Mbps both upstream and downstream. data rate of 1.5 Mbps and telcos still had some trouble conditioning the While VDSL2 works on twisted-pair, it has to be very good twisted-pair. The twisted pair to handle it. VDSL2, as deployed in the U.S., has data rates of challenge that telcos face is to ensure that the twisted pair in their system is approximately 50 Mbps, or, about thirty times faster than ADSL1. capable of supporting VDSL2 service.

2 Guide to the Last Mile www.megger.com Guide to the Last Mile 3 Telcos have been successful with these older services by having a reactive for voice-grade POTS. When a tech investigates trouble on a circuit with a approach to maintenance. In simple terms, they try to install the service DSL service, he or she must make sure that the copper, the physical layer, is and if it doesn’t work or if the subscriber complains, they look for another in excellent condition before testing the service. If there is a physical-layer pair to use. As a last resort, they try to fix the problem. Now they have to problem, it doesn’t matter what condition the DSL equipment is in, because become proactive when they implement VDSL for IPTV. As we’ll discuss the service won’t work properly. Once the technician is certain that the below, with the older services, a subscriber may not even be aware that the physical-layer is in good shape, he or she then goes on to test the service. service is substandard. However, with real-time, very high speed services, the subscriber is aware of the smallest degradation of service. Consequently, The second step is to verify that there is Layer 2 connectivity. Does the the pair should be thoroughly tested (qualified) before this real-time, very (subscriber premises) DSL modem work? Will it communicate with the high speed service is installed on it. The tech or the telco can’t merely install (remote cabinet) DSLAM? Will they “sync”? This is accomplished by it on a likely pair and, if the subscriber complains, switch pairs and hope replacing, in turn, the modem and the DSLAM with known-good (“golden”) that the subscriber doesn’t complain again. If the tech finds a fault when modem and DSLAM and trying to sync. The tech also verifies that these conditioning a pair for VDSL2, he or she doesn’t merely look for another devices are configured properly. pair. Rather, the tech finds and removes the fault. The next step is to verify that there is Layer 3 connectivity to and through the Internet. Does the BRAS connect to other BRASs in the network? Will Basic Testing Procedure it “surf”? The tech can determine this by trying to ping an IP address of a What are the elements of the VDSL network that we need to test? Basically, known server that is “on the other side” of the internet. If this is successful, VDSL works by having a DSL modem or, for “Triple Play,” a DSL router at the the tech is said to be able to “surf.” customer’s premise connect to and communicate with a digital subscriber line access multiplexer (DSLAM) at the telco’s central office or, more likely, The key to testing is to make sure it is done one layer at a time and that at a remote cabinet that is connected to the central office via fiber. From the technician works his or her way from the lowest layer (physical layer) our discussion above of the 7-Layer protocol stack, this DSLAM is a Layer 2 upward. If a circuit won’t surf, the tech does not start the trouble-shooting device. Layer 1 (the copper physical layer) must be in good condition for the process by testing Layer 3 connectivity (trying to ping an IP address across the Layer 2 DSLAM to work and the DSLAM must work before Layer 3 can work. Internet). After all, it might not surf because it doesn’t sync. A DSLAM is a large group of modems that combines the Ethernet signals received from the subscriber modems and multiplexes them into a single, in this case, Internet Protocol (IP) network and transmits this to the Broadband Remote Access Server (BRAS).

The BRAS is a Layer 3 device located at the center of the telco’s IP network and combines the traffic from many DSLAMs. It routes this IP traffic to other BRASs in the telco’s network or outward in the Internet to Internet routers. It is the BRAS that assigns the subscriber’s IP address.

We’ll cover specifics later but to outline the basics of the testing process for a VDSL2 circuit that does not work at all, the first step is physical-layer testing which determines whether or not the Layer 1 (twisted-pair copper OSP) is capable of handling a very high-speed digital service such as IPTV. The pair must be more rigorously tested than if it merely were going to be used Figure 1: VDSL2 network elements

4 Guide to the Last Mile www.megger.com Guide to the Last Mile 5 Characteristics of a Perfect Pair for Installing VDSL2 Service There should be no impulse noise hits above -35 dBm on the pair. This is Measured wire-to-wire (tip-to-ring or A-to-B) and each wire to the cable’s measured across the spectrum used by VDSL2 during a five-minute test. shield (ground), a pair should have no DCV but ≤10 DCV is acceptable. If it Impulse noise spikes cause pixilization errors and, when there are enough has >10 DCV, it could be crossed with a working pair. of them, dropouts.

Measured wire-to wire (tip-to-ring or A-to-B) and each wire to the cable’s VDSL2, as implemented following ITU G.993.2, has length limits. Data shield (ground), a pair should have no ACV but ≤10 ACV is acceptable. If it throughput of 50 Mbps can be achieved over a distance of 3,300 ft (1,000 has >10 ACV, it might be dangerous and it is a source of noise. meters) on a good copper pair and data throughput of 30 Mbps can be achieved at a distance of 3,500 ft (1,066 meters) on a good copper pair. Measured wire-to wire (tip-to-ring or A-to-B) and each wire to the cable’s shield (ground), the resistance should be close to the test set’s reading of A pair cannot have bridged taps because they cause a decrease in data infinity but >3 ΩM is acceptable. The pair has a fault if the wires are shorted throughput. to each other or one of the wires is shorted to the shield by ≤3 MΩ because There should be a very minimum of impedance changes along the pair. it is unbalanced and is susceptible to noise. It is also susceptible to corrosion. Impedance changes cause some of the signal to be bounced back toward A pair cannot have any load coils because they kill any signal above 4 KHz the transmitter. This adversely affects the length over which the signal is such as VDSL. intelligible.

A pair should be balanced. Put simply, tip (A) should be a mirror image of Factors That Can Cause a Pair to be Ineffective for VDSL Service ring (B). If it is not, the pair will be susceptible to noise which will reduce the There are several characteristics that would cause a pair to not qualify for data throughput, sometimes to the point of making the signal unintelligible. VDSL2. The noise on the pair, measured across the spectrum used by the VDSL2 being tested (8, 12, 17 or 30 MHz) should be ≤-60 dBm. Noise above this Breakdown of insulation resistance power will cause diminished data throughput or, in some cases, loss of signal. As stated elsewhere here, the insulation on the conductors has to be good enough to prevent high-resistance faults. Insulation resistance can be Telcos sometimes state power levels in “decibels above reference noise” compromised by: (dBrn) instead of “decibel above one milliwatt” (dBm). Using dBrn avoids the use of negative numbers. If noise specifications are stated this way, the n Being pierced with even a bed-of-nails connector can allow moisture to conversion is: penetrate to the copper which causes corrosion and high-resistance faults. 0 dBm = 90 dBrn n Water in the cable will cause not only a breakdown of insulation -90 dBm = 0 dBrn resistance but also changes the impedance of the pairs which causes reach For measuring the power of noise and power influence on POTS (voice-band and data throughput to be adversely affected. or <4 KHz) circuits, telcos in the U.S. further refine dBrn by using a filter n Not being careful when handling the pairs can cause the insulation of that mimics what the human ear can hear. As is discussed elsewhere, they the conductors to be degraded. It can get nicked, abraded and otherwise don’t want to measure noise that does not adversely affect the ability of damaged. The advantage of going to “ready access plant” as telcos did the receiver to understand the information that was transmitted. This filter years ago is that any tech can access it. The disadvantage is that any tech is called a “C Message” or “C” filter. The power level of noise so filtered is can access it. It is not uncommon for technicians to be a source of trouble measured in “dBrnC”. Places other than the USA use a slightly different filter in the cable plant. to accomplish this by using what is called a psophometric filter.

6 Guide to the Last Mile www.megger.com Guide to the Last Mile 7 Grounding A properly bonded cable shield keeps noise away from the cable pairs. Good grounds deteriorate over time allowing noise that would have How does it do that? This properly bonded shield requires a low resistance otherwise gone to ground to enter the pairs. conduction path from one end of a section of shield back to the other end of that section through a path outside the cable. This connection effectively n The actual grounding conductors (usually ground rods) can degrade by shorts out voltage on the cable shield and allows a current to flow along the corroding, coming loose in the earth, or from tampering. shield that is induced by magnetic fields from a near-by power cable or other n The ability of the earth to conduct (resistivity) can change as the soil dries noise sources. That induced current creates a magnetic field that opposes the out such that a once-adequate ground in the wet season is no longer magnetic field coming from power, this cancelling out longitudinal voltages adequate in the dry season. Changes in topography such as grading, etc. on the cable pairs. can also cause this. A ground rod bonded to the shield at each end of a cable section creates a n It is even possible that an adequate ground was removed either shorting path between the ends of that shield. But, resistance on the ground accidentally or in a misguided attempt to reduce noise. rods can allow voltage to be present between the ends of the section. This voltage appears on the pairs that are in the cable. Consequently, a very n Ground rods at the subscriber’s location: the station protector or network low-resistance ground is required at each end of the section which is usually interface device (NID) are part of the protection against lightening and provided by bonding the shield to the power MGN. Since induced power power surges. They are tested to assure that they provide adequate safety varies along the length of the cable, it is important to divide the cable into protection. A maximum resistance value of 25Ω is considered low enough shorter sections so that the shield is bonded to the MGN at least every 1,000 to assure safety. However, adequate noise reduction is not achieved ft (300 meters) and also at each point where the exposure to power changes, unless that ground rod is bonded to the incoming power neutral. Also, such as where it branches. safety is not assured unless the incoming power neutral is bonded to the subscriber’s ground (station ground). Dangerous voltages can be present Properly installed bonds deteriorate because: on the station ground if it is not so bonded to the incoming power n Over time mechanical bonds may loosen due to thermal changes and neutral. corrosion. n Ground rods along a cable path, bonded to the cable shield, are effective n Bonds are removed in order to use a cable locator. Sometimes they are not for lightning protection, but frequently do not provide a low enough replaced. resistance path to be effective in reducing power influence or broadband noise (both discussed below) on cable pairs. This creates the need to n Bonds may be removed to check ground resistance. Sometimes they are periodically bond the cable’s shield to the power company’s multi-ground not replaced. neutral (MGN) as described below. n Bonds may be removed accidentally or in a misguided attempt to reduce noise. Bonding The bonds that electrically connect the cable shields together at a splice and may also connect the shields to a ground rod or to the power company’s Bridged tap MGN can be degraded which, again, allows noise to enter the pairs that A section of cable pair that has been added between the “normal” ends of would otherwise have gone to ground. (Bonds can be installed incorrectly or the pair is known as a bridged tap. Another pair has been “T”ed onto the not at all but the assumption here is that they were adequate at one time.) original pair. This is very common and, in fact, it was standard operating

8 Guide to the Last Mile www.megger.com Guide to the Last Mile 9 practice to use them in POTS. The dilemma is that that they cause problems because of the impedance change, these pairs are not only susceptible to with services such as DSL. In VDSL2, bridged taps as short as 3 ft (1 m) reduce crosstalk; they are a source of crosstalk in other pairs. When we think of data throughput. crosstalk, we usually think of it as an audible signal in the voice frequency range. This crosstalk, however, is at VDSL frequencies. Telcos have Load coils programs to retrofit these cabinets with data-grade wire. On copper circuits in excess of 18,000 feet, capacitance of the pair causes high loss to voice frequency signals. The solution was to put load coils on Unbalanced pair these long circuits in standard locations along their length. Load coils are This occurs when the tip (A) conductor is not a mirror image of the ring (B) series inductors that reduce capacitance loss up to 4 KHz (voice band) but conductor. It makes the pair intermittently susceptible to noise which inhibits that also kill frequencies above 4 KHz. This was great for POTS circuits but it data throughput. It is caused by: kills ISDN and DSL. While load coils are rarely installed today, there are LOTS of them left out there and they must be removed before any of the above- Tip and ring (A and B) are not the same length which might, at first glance, mentioned services will work. seem improbable. After all, both conductors go from the transmitter to the subscriber. Consider an unbalanced bridged tap. This is where one of the conductors of the bridged tap has been removed but the other one was not Excess length removed. VDSL2 data throughput deteriorates quickly from a theoretical maximum of 100 Mbps at the source to about 50 Mbps at 3,300 ft (1,000 meters) and to Tip and ring (A and B) are not the same resistance which is caused by a about 30 Mbps at 3,500 ft (1,066 meters) over a good pair. high resistance fault on one of the conductors. The insulation of only one conductor may have been nicked which, over time, allowed it to corrode. Impedance changes Or, some corrosion occurred on one conductor at a splice point where the Too many changes in impedance cause loss of range and data throughput connector was improperly crimped. problems. In addition to the changes caused by wet cable mentioned above, impedance changes are also caused by anything that changes the distance The pair is being recombined. This is a splicing error where one of the between the tip and ring (A and B) conductors such as: pair’s conductors, e.g. the tip (A), was spliced to the tip (A) of another pair, allowed to go on some distance and then spliced back to the original pair. n Single-wire splice connectors such as “B” connectors or Scotchlok® that, Not only is there a length difference because of the different twist ratios of rather than splicing the wires of a twenty-five-pair group at once, would the two pairs, but there is an impedance mismatch. almost invite the tech to untwist the pair because he or she had to splice one wire of the pair at a time. The tech may have tried to squeeze the two Power company wires of the pair together when the splice was complete. The tech may or Even with good grounding and bonding, a problem with the power may not have re-twisted them and, if he or she did, it is not likely that the company having a phase imbalance, inadequate grounding or a faulty twist was the same as in the rest of the pair. transformer can cause so much induction that the circuit will be noisy. n The connecting wires in a cross connect are the pairs that connect the two sides of a cross connect cabinet which might be untwisted or not sufficiently twisted. In addition to causing data throughput problems

10 Guide to the Last Mile www.megger.com Guide to the Last Mile 11 Tests to Effectively Deliver VDSL Service n The conductors of the pair (tip and ring or A and B) would not be shorted With a volt/ohmmeter, the following measurements are made: to the cable’s shield or to any other ground. n Wire to wire – tip-to-ring or A-to-B As a practical matter, however, resistance readings >3 MΩ are acceptable. n Each wire to shield - tip-to-shield [ground] and ring-to-shield [ground] or, There are two ways to measure insulation resistance; the first is with a A-to-shield [earth] and B-to-shield [earth] standard low resistance ohmmeter (found in multimeters). RESISTANCE AC volts This is a major cause of noise. There should be almost no AC voltage on any +999 MΩ cable pair that is not ringing. (Ringing voltage is between 70 ACV and 110 ACV.) If there is more than about 10 ACV on a line when it is not in ringing +999 MΩ mode, check for bonding problems on the cable shield and for ground problems. In rare cases the pair may be crossed with AC. In any case, the AC +999 MΩ needs to be removed. (Some telco-specific volt/ohmmeters provide a user- selectable option of a 100 KΩ or a 3 MΩ termination. Select 3 MΩ.) REV

DC volts Figure 3. Sample of resistance test display A non-working pair ideally should have no DCV. If it has over 10 DCV, it could be crossed with a working pair. The second way to measure insulation resistance is by a “leakage test.” This tests insulation resistance at 150 DCV which is higher than the normal

operating voltages in order to determine the likelihood of a future breakdown in insulation (a high resistance fault) and/or to bring about such an impending failure while the tech is on site and can locate and repair the problem.

LEAKAGE

Figure 2. Sample of volt/ohmmeter test display 1.0 Mohm

Resistance The insulation on the wires of the pair that is being qualified for VDSL REV ideally would have no shorts; the tip-to-ring (A-to-B) and tip-or-ring-to-shield (A-or-B-to-shield) resistance would be infinite. Figure 4. Sample of leakage test display n Tip-to-ground (A-to-earth) and ring-to-ground (B-to-earth) would be As with the volt/ohmmeter insulation resistance test discussed above, if there completely open. is ≤3 MΩ insulation resistance, there is a fault.

12 Guide to the Last Mile www.megger.com Guide to the Last Mile 13 While these insulation resistance tests will determine if there is a high- Figure 6 shows a result screen when the RFL test is complete. The resistance fault, they will not locate it. A resistive fault locator will locate the abbreviations are: fault. DTS – distance to strap (from the test set to the other end) DTF – distance to fault (from the test set) Resistive fault locator (RFL) FTS – Fault to strap A multimeter can find hard, or “metallic”, faults. These are dead shorts and wide opens. Because, at a given temperature, the resistance per foot of each The screen in Figure 6 shows the fault-to-strap distance was calculated by gauge of wire is known, distance to hard faults can be calculated. subtracting distance-to-fault from distance-to-strap. This is done when only one good conductor is used. If a pair of good wires (instead of a single wire) Other faults, resistive faults, are high resistance. Metal does not touch metal. had been used, the fault-to-strap distance is actually measured electrically These resistive faults make up the majority of faults in OSP. They are what and the display would indicate the result was “verified.” This latter method we are looking for with the leakage test; an insulation resistance test. While allows the tech to know whether the readings make sense: does distance-to- the leakage test can indicate that we have one of these high resistance fault added to fault-to-strap equal distance-to-strap? faults and its severity, it doesn’t help us locate it. A TDR won’t find it either, however, an RFL will find faults of <2 ΩM on a pair or a single conductor. RFL

If a tech has at least one good wire (preferably a pair of good wires) in the DTS 374 FT same cable as is the faulted pair, he or she can determine the location of the fault. Because resistance of metal varies with the temperature and the size 204 FT 170 FT DTF FTScalc of the conductor, the tech must know the gauge of the conductors and the temperature of the cable. The test requires that a strap be used at the end SIZE 0.75 MΩ SETTINGS of the cable away from the test set. Modern RFL sets have diagrams showing GAUGE = 24 TEMP = 70 F the tech how to set up and perform the test as shown in Figure 5. EXIT START SETUP

Yellow GOOD WIRE2 (OPTIONAL)

Green GOOD WIRE1 Figure 6. Sample of completed RFL test display

Red FAULTED WIRE1 STRAP Load coil detection Black FAULTED WIRE2 (OR SHIELD) The requirement for a DSL circuit is that it have no load coils. Presence of HT1000 ACV or DCV on the pair can inhibit the ability of a test set to detect the CONNECT THE HT1000 AND STRAP TO FAULTED CABLE ACCORDING TO FIGURE correct number of load coils so, if relevant, it is recommended that central PRESS ENTER TO CONTINUE office battery be removed before testing.

A load-coil-detecting function in a test set display will show the number of Figure 5. Sample of RFL test display load coils detected. A waveform graph also may be displayed. The number of dips in this waveform corresponds to the number of load coils. The waveform does not, however, indicate the distance to the load coils. A TDR will give the distance to the first load coil (covered in the next section).

14 Guide to the Last Mile www.megger.com Guide to the Last Mile 15 NOISE COILS 3 46.2 dB

PRESS RTN TO EXIT USE KEYPAD TO DIAL START LOSS NOISE PWR INF DIAL w/c FIL LIST

Figure 7. Sample of load coil test display Figure 8. Sample of noise test display

Balance Voice-band power influence The pair should be balanced. Tip (A) should be a mirror image of ring (B). Power influence is the noise that is caused by AC induction on a cable or pair. The tech determines the extent to which a pair is balanced by looking at the If the cable’s shield is bonded together and periodically bonded to ground, symptom of imbalance; the level of noise on the pair. We will consider: this induced AC goes to ground and does not find its way onto tip and ring n Voice-band noise (A and B). To perform the power influence test, tip and ring (A and B) are shorted and the noise is measured to the shield. In a “balanced pair,” tip and n Voice-band power influence ring (A and B) are mirror images of each other. In such a balanced pair, the noise due to power influence is not audible. In an unbalanced pair, the noise n Longitudinal balance is audible. Acceptable power influence readings for VDSL are≤ 90 dBrnC. n Stress Unacceptable levels of power influence are typically caused by poor bonding and grounding of the cable shield. While the need to measure the first three is largely replaced by the stress/ super stress test, we still need to understand them. POWER INFLUENCE

Voice-band noise From the field, a tech connects to a quiet termination line, (by dialing a special number) in the central office, that produces dead silence. Noise on 80.9 dB the circuit is then measured between tip and ring (A and B). Any noise that the tech measures has to be from the OSP portion of the circuit under test. PRESS PI BUTTON TO INS/REMOVE C FILTER Since vacant pairs are not connected to the network, this measurement LOSS NOISE PWR INF DIAL w/c FIL LIST cannot be made on vacant pairs because of the need to have a working pair to access a quiet termination number. Passing noise measurements are ≤60 dBm. Figure 9. Sample of power influence test display

16 Guide to the Last Mile www.megger.com Guide to the Last Mile 17 Longitudinal balance Readings for stress vary from telco to telco so the following thresholds are a As previously discussed, if tip and ring (A and B) of a cable are not alike (if guideline: they are “unbalanced”), the pair is susceptible to having noise induced onto it. Telco engineers have, therefore, long been on a quest to be able to easily Good: < 20 dBrn determine how balanced a cable pair is; how alike tip and ring (A and B) of Marginal: 20 to 30 dBrn that cable pair are. Longitudinal balance, dating from the 1970s, gets at this. It is a calculated number arrived at by measuring power influence (discussed Bad: > 30 dBrn above) on a circuit and then measuring circuit noise (discussed above) on If the pair fails the stress test, it is unbalanced. Tip (A) is not a mirror image the same circuit. Circuit noise is subtracted from power influence. A passing of ring (B). Reasons why the pair fails include: longitudinal balance number is ≥60 dB or as high as possible while passing the power influence and noise tests. n One conductor of the pair may be longer than the other. If the open meter test (discussed below) were performed, tip-to-shield (A-to-shield) But, the point is that longitudinal balance does not really give an length would be different than ring-to-shield (B-to-shield) length. indication of how alike or how balanced tip and ring (A and B) are. They merely provide an indication of how noisy a pair is only at the time the n A split pair will cause the pair to be unbalanced. A TDR will show where measurements are made. the split is. Use a tone and probe to identify the actual wires involved. Listen for tone carry-over on adjacent conductors. The tone on the split Longitudinal balance varies by time of day and season of the year because conductor will be louder than the tone on adjacent conductors. power influence does. Therefore, a tech might get a passing longitudinal balance number one time and a failing one another time. Longitudinal n A series resistance fault in tip or ring (A or B) will cause an imbalance. balance tests are not repeatable. An unbalanced pair may only be noisy Short and ground the pair at the far end. Using a multimeter test, measure when the garage door goes up or when the air conditioner is running while the resistance of each side to ground. The highest reading is the series the washing machine is running or when a light with a bad ballast is turned fault. The fault can be located using a TDR (discussed above). on. If the tech doesn’t do the longitudinal balance test while that condition exists, the circuit will not fail the test. n A short between the conductors is a fault but it does not cause an imbalance because tip and ring (A and B) are equally affected; they are Stress test still mirror images of each other. This short will not cause a stress/super Very simply, the stress test puts a known constant amount of power stress test to fail. These faults can be located with a TDR. influence on the pair and then measures the resultant noise. Stress provides n One of the conductors being shorted to the shield (called a “ground”) will, a very good indication of the extent to which tip and ring (A and B) are however, cause an imbalance which will cause the stress/super stress test to mirror images of each other; the extent to which they are balanced. fail. An RFL will locate a ground. Stress overcomes the disadvantages of the longitudinal balance test: Noise n The stress test results do not vary when power influence changes. It is The noise on the pair, measured across the spectrum used by VDSL2 (up to repeatable. 8, 12, 17 or 30 MHz), should be <-60 dBm. We have discussed noise as the symptom of an unbalanced pair. There are some other helpful ways to look n The stress test can be used on vacant pairs because there is no need to dial at noise. But, there are several questions to consider when tackling noise. into a quiet termination line to use it.

18 Guide to the Last Mile www.megger.com Guide to the Last Mile 19 n Is it relevant? Is its frequency such that it will interfere with the signal? As we’ve said, a good circuit will have ≤30dBrnC of noise. If there is high NM, the spectrum analyzer is used to look at the frequency of the noise to see n How powerful is it? Is it powerful enough to interfere with the signal? if it is from PI (60 Hz and its harmonics) or from something else; something n Is it constant or intermittent? If it is intermittent, is it extremely short that is coming from within the cable itself such as crosstalk. Remember that duration spikes? NM is between tip and ring (A and B). It is inside the cable. n Where is it coming from? How do we get rid of it? If the pair being measured is well-balanced, the NM should be low no matter how high PI is. However, if there is high PI (on this well-balanced pair) and Voice-band spectrum analyzer low noise, there will be problems with VDSL2 because the high PI number This is the first test tool we’ll consider. When noise or power influence indicates problems with grounds. This means that there is likely to be is measured, the result is a reading of the power (dBm) of that noise. impulse noise (more later). In trying to find the source of that noise, knowing its power level is not The tech can compare information from both spectrum analyzer graphs much help. The tech needs to analyze that noise to find its frequency in (Power Influence and Noise Metallic) to determine the cause of noise order to determine whether the noise comes from AC and/or its harmonics problems. (outside the cable) or from crosstalk (inside the cable). That is what a spectrum analyzer does. It is a graphical representation of the noise with A voice-band spectrum analyzer screen displaying approximately a -45 dBm the frequency of the noise on the horizontal axis and the amplitude of these noise source at 1,344.1 Hz is shown in Figure 10. frequencies on the vertical axis. POWER SPECTRUM 4162.0 Hz Before further discussing the spectrum analyzer, let’s review the two ways dBm 0 of measuring noise from above. The first is “Power Influence” (PI). This is -10 measured by shorting tip and ring (A and B) and measuring them to ground -20 or “longitudinally”. The second is “Noise Metallic” (NM) or simply “Noise”. -30 It is measured between tip and ring (A and B) and is done on a voice-band -40 -50 spectrum analyzer by connecting the analyzer’s leads to tip and ring -60 (A and B) of the pair. -25 dBm 1344.1 Hz When there are high power influence (PI) readings, the cause is almost certainly bad grounds and bonds or noise from the commercial power Figure 10. Sample of voice-band spectrum analyzer display ground getting into the telco ground. A telco practice of tying the cross- connect cabinet’s ground to the power-pole MGN might cause high power Wide-band spectrum analyzer influence if the MGN is open, causing power neutral current to travel back A wide-band spectrum analyzer is used much like a voice-band spectrum toward the substation on the telco neutral.. If there is high PI, the tech analyzer. The latter is useful in determining if commercial AC noise is present looks at the power influence spectrum analyzer to verify that it is 60 Hz or but it doesn’t show noise across the frequency band occupied by VDSL. The a harmonic. This is done on the analyzer by connecting the black lead to tip wide-band spectrum does that. (Unlike ADSL2, VDSL2 does not have a fixed (A) and the red lead to the cable shield and leaving the green lead, if the bandwidth. VDSL2 has different “band plans”. The bandwidth can be 8, 12 analyzer has one, disconnected. 17 or 30 MHz. Even within these bandwidths, there are different versions [8a, 8b, 12a…].) The point is that the tech wants to look for noise within the frequencies relevant to the VDSL2 that he or she is working with.

20 Guide to the Last Mile www.megger.com Guide to the Last Mile 21 The shape of that noise on the wide-band spectrum analyzer can help With telcos introducing very-high-speed real-time services such as IPTV, identify its source. In general, if the noise is spread widely across the impulse noise becomes very important for two reasons. The first is that spectrum, it is likely from crosstalk which is from inside the cable. If there these services have very small bit times so a relatively small impulse-noise are spikes or if the noise is concentrated at the low end of the spectrum, it is spike would knock out more data. The second is that these services are real- from outside the cable and is caused by bonding and grounding problems. time. In other words, the data cannot be retransmitted so the effect is not masked by merely having the effective data rate go down. The subscriber A wide-band spectrum analyzer showing noise up to 33.2 MHz with about a immediately sees the effect of the impulse noise spike in the form of -32 dBm noise spike at 1.42 MHz and a noise floor at about -75 dBm appears pixelization or even picture dropout. Sometimes the transmission equipment in Figure 11. even has to “reset” itself by retraining. SPECTRUM 33.2 MHz dBm -20 Background -30 These very-high-speed services use modern video protocols that “compress” -40 -50 data to get more and more video channels out of the same bandwidth. -60 To perhaps oversimplify it, they make every bit carry more and more -70 -80 information. This makes the loss of a few bits have a disproportionately large -90 -20 dBm1.42 MHz impact.

LESS MORE VBAND WBAND FREQ FREQ TONE TONE While a growth industry recently has been in mitigating the effects of impulse noise, telcos have known for over a hundred years that shielding Figure 11. Sample of wide-band spectrum analyzer display their cables and grounding these shields will eliminate noise on a balanced pair. Some telephone people are relearning this fact. Impulse noise When we were outlining the questions with respect to noise above, one of What is an impulse noise test? the questions dealt with whether the noise was extremely short duration An impulse noise test detects and counts noise spikes or “hits” over a period spikes. These very short duration spikes of noise are called “impulse noise.” of time and displays the results numerically This impulse noise poses a special threat to a digital communication system. If a spike of noise could make the receiver of such a system not understand How impulse noise is measured whether a plus or a minus had been transmitted, the whole string of digital There are several variables for the tech to determine and, in some cases, set information might have to be retransmitted. There are algorithms that when measuring impulse noise. Among them are: can correct one or two such errors. But as bit rates increase, the “time” or “space” of a single bit gets smaller and smaller and the same size impulse Measurement filter noise spike can knock out more information. This makes it more likely the When we want to measure the noise that affects what our ears hear, we whole block of information will have to be retransmitted. In the world of use a filter that behaves as our ears do; one that does not pass anything data communications technology such as ADSL, the result of this is that the below about 400 Hz or above 4,000 Hz. Basically, it rejects noise that our subscriber merely sees a slowdown in effective data rate. ears cannot hear. We would not want to say that an audio circuit was bad While impulse noise caused problems to data services such as ADSL, that because it had noise in the 50 KHz range; noise that we cannot hear. So, we service is structured so that such problems only cause the effective data disregard noise that we cannot hear. When we measure impulse noise we are rate to be reduced. Consequently, the subscriber may not be aware of the problem.

22 Guide to the Last Mile www.megger.com Guide to the Last Mile 23 concerned only with noise that will adversely affect the service on the circuit under test. So, test sets that measure impulse noise use different filters to IMP NOISE reject this noise. 5 TOT HITS “F” – for HDSL – from 5 KHz to 245 KHz 29 SECONDS “G” - for ADSL – from 20 KHz to 1.2 MHz 7 1 HITS 3dB LOWER HITS 3dB HIGHER SETTINGS “J” – for ADSL2 – from 20 KHz to 2.2 MHz FILTER = G LEVEL = -45 dBm

RESTART SETUP “Max” – for VDSL2 – from 20 KHz to 30 MHz

Test time Figure 12. Sample of impulse noise test screen display How long do we want the test to run? Back in “the day” (‘70s and ’80s) impulse noise tests for such things as radio remote-broadcast circuits and Length measurement circuits that carried TV broadcasts for the networks were run for hours Data throughput on VDSL2 of 50 Mbps can be achieved over a distance of and even days. After all, impulse noise varies over time like other power- 3,300 ft (1,000 meters) on a good copper pair and data throughput of 30 influence noise. A common standard for VDSL now is that there are no hits Mbps can be achieved at a distance of 3,500 ft (1,067 meters) on a good during a 5 minute test. copper pair. So, there is a limit on the length of a pair being qualified for VDSL2. Trigger level Impulse noise test sets automatically set power thresholds for impulse noise Length of a cable pair can be measured electrically in three ways: detection. This is called the “trigger level.” Some test sets have additional n Resistance – Discussed above in “RFL” section trigger levels at both 3 dB above and below the “main” trigger level and then separately count hits that reach all three of these. This allows the n Impedance – Discussed later in “TDR” section impulse noise hits to be “qualified.” A threshold of -35 dBm is generally accepted. n Capacitance – Discussed below in “Open Meter” section

There should be no impulse noise hits during a five-minute test. If the pair Open meter fails the impulse noise test, the likely culprit is bad grounding or bonding A capacitor is two conductors separated by an insulator. So, a pair of wires, that is allowing the noise spikes into the pair instead of taking them to as well as one wire and the cable shield, are both capacitors. In other words, ground. there are three conductors: tip (A), ring (B) and shield. Therefore, there are three capacitors: Figure 12 shows a sample of an impulse-noise-test screen for a test with a “G” filter (20 KHz to 1.2 MHz) that has been running for 29 seconds and has n Tip-to-ring (A-to-B) recorded 5 hits at the main trigger level of -45 dBm with 7 hits at -48 dBm and 1 at (the noisier) -42 dBm. n Tip-to shield (A-to-shield)

n Ring-to shield (B-to-shield)

24 Guide to the Last Mile www.megger.com Guide to the Last Mile 25 The capacitance of OSP cable in the USA is engineered to have values of: Therefore, the tip-to-shield (A-to-shield) length and the ring-to-shield (B-to- n Tip-to-ring = 0.083 µF/mile (microfarads per mile) or, A-to-B = 0.0516 µF/km shield) length must be almost equal and less than 3,300 ft (1,000 meters) (microfarads per kilometer) for a data throughput of 50 Mbps and less than 3,500 ft (1,066 meters) for a data throughput of 30 Mbps. n Tip-to-shield and ring-to-shield = 0.125 µF/mile or, A-to-shield and B-to-shield = 0.078 µF/km <0 ft Gain = 0 893 ft> <0 ft Gain = 0 893 ft> OPEN METER 904 FT 859 FT

901 FT VOP=0.670 519.1 ft VOP=0.670 519.1 ft Cursor Cursor OPEN CAPperDIST uFperMi TR = 0.083 XG = 0.135 LESS MORE LESS MORE CABLE CABLE ZOOM SETUP CABLE CABLE ZOOM SETUP T-R T-G R-G

Figure 14. Figure 15.

Figure 13. Sample of open meter screen display Time Domain Reflectometer (TDR) (Note that drop wire, inside wire, etc. is not engineered to have this The impedance of a pair changes when the distance between tip and capacitance and their length cannot be measured with an open meter unless ring (A and B) changes or when the material (air, water, jelly) that fills the their capacitance per foot (meter) is known and the open meter allows that cable changes. A TDR detects these changes in impedance and, without capacitance to be input.) going deeply into how, gives the distance to the impedance changes. Very simply put, a TDR operates by sending pulses of energy down the wires and The amount of capacitance of two conductors is proportional to their length. timing how long it takes for them to get back after being reflected by an An open meter simply measures the capacitance and does the arithmetic impedance change. The TDR knows how fast these pulses travel because the to convert that to length. It is able to measure the length of any one of tech has input that speed. By knowing how long a pulse has been gone and these conductors against any other. The tech does not need to know the how fast it is moving, the distance covered by the pulse can be calculated. A temperature or gauge of the wire or how fast an electrical signal moves simple rule for interpreting a TDR trace is that the impedance and the trace along the pair so the open meter is easier to use for length measurements go up as tip and ring (A and B) move apart, as seen in Figure 14, and they go than are the other two methods discussed that are based on resistance and down as tip and ring (A and B) move closer together, as seen in Figure 15. In impedance. order to qualify the last mile of copper for VDSL2, a tech isn’t using a TDR to locate shorts, opens, splits and wet cable at long distances. That tech is now The tech is not so much interested in tip-to-ring (A-to-B) length as he or she looking for changes in impedance that didn’t cause problems for POTS. And, is in tip-to-shield (A-to-shield) length and ring-to-shield (B-to-shield) length while those “old” faults do cause the impedance to change, the tech is now and making sure that they are the same. If they are not, the shorter one is looking for different kinds of “faults” that also cause a change in impedance open. (The shield goes as far as the longest wire does.) and a reduction in data throughput in VDSL2. The tech is now looking for the following conditions:

26 Guide to the Last Mile www.megger.com Guide to the Last Mile 27 Bridged tap The connecting wire in a cross connect In the past, long-range TDRs were the ones that sold most because they These are easy to locate by opening the cabinet and visually inspecting the helped locate the faults that caused problems for the OSP of that day; opens, wires. On a TDR trace, however, they may be dscribed as: tip and ring (A and shorts and wet cable at a relatively long distance away. Those TDRs did show B) are separated, increasing the impedance which causes the trace to go bridged taps but because of the nature of the TDR-trace dead-zone, short upward a bit but not as high as the complete open at the end of the pair. ones were masked by the dead zone that follows the splice that created the bridged tap. A TDR for conditioning twisted-pair OSP for VDSL would need a Damp or wet pairs very short pulse width so as to minimize the dead zone. Moisture decreases the impedance causing the TDR trace to go downward as illustrated so the beginning of the wet section can be located. But, as it also

<0 ft Gain = 0 893 ft> changes the speed with which the TDR’s pulses of energy travel on the wires, the length of the wet section cannot be determined unless the tech goes to the other end and tests from that direction. In Figure 17, the wet section is between the two cursors.

<0 ft Gain = 0 893 ft> VOP=0.670 133.5 272.9 ft Delta Cursor

LESS MORE CABLE CABLE ZOOM SETUP

Figure 16. Sample of bridged tap VOP=0.670 169.1 320.0 ft Delta Cursor

A TDR trace of a bridged tap is shown in Figure 16. The location of the splice LESS MORE CABLE CABLE ZOOM SETUP that makes the bridged tap is marked by the downward movement of the trace. The next upward movement of the trace is considered to be the end Figure 17. Sample of wet section of the tap. A pair with an open at the end of the bridged tap and an open at the end of the main part of the pair would have two of these upward movements of the trace. This is the only time a TDR can see two actual opens Split and re-split pair on a single pair. Figure 18 shows the trace for a recombined pair. The trace goes up at the first cursor which is the point where the pair was split because the Single-wire splice connectors impedance goes up when the wires are separated. The trace then goes down These may not actually need to be located but if they are, the TDR trace at the second cursor which is the point of the re-split where the wires are may be described as: tip and ring (A and B) are untwisted to begin the splice rejoined. so the trace goes up and then immediately down because the wires are brought together to be inserted into the connector. The height of the trace can be manipulated by changing the gain and pulse width. However, a good splice, one with very little impedance change, will be harder to find than one with a larger impedance change.

28 Guide to the Last Mile www.megger.com Guide to the Last Mile 29 Step 2 <0 ft Gain = 0 893 ft> Using the auto-test capability of some VDSL test sets, test the physical copper pair (Layer 1) for the following:

n DC volts tip-to-ring (A-to-B) and tip/ring to-ground (A/B to-earth) = <.5 DCV is ideal but ≤10 DCV is acceptable.

VOP=0.670 118.7 255.1 ft Delta Cursor n AC volts tip-to-ring (A-to-B) and tip/ring-to-ground (A/B to-earth) = <1.0 LESS MORE ACV is ideal but ≤10 ACV is acceptable. CABLE CABLE ZOOM SETUP

n Stress = <20 dBrn is good, 20 to 30 dBrn is marginal.

Figure 18. Sample of recombined pair n Leakage tip-to-ring (A-to-B) and tip/ring-to-ground (A/B to-earth) = Since most TDR traces don’t look like illustrations in user manuals with the infinity is ideal [no leakage] but >3 ΩM is acceptable. horizontal part of the trace being nice and flat before they change to a n Resistance tip-to-ring (A-to-B) and tip/ring-to-ground (A/B to-earth) = well-defined event, a dual-trace TDR allows the comparison of the trace of a infinity is ideal [no leakage] but >3 ΩM is acceptable. good pair with that of a problem pair by showing the differences in the two fairly bumpy traces. n Open meter tip-to-ring (A-to-B) and tip/ring-to-shield (A/B to-earth) = <3,300 ft (1,000 m) for 50 Mbps date throughput and <3,500 ft (1,066 m) Steps to Turn Up a VDSL Circuit for 30 Mbps data throughput. Step 1 n Load coils = none. Identify the VDSL circuit appearance at the DSLAM and note the upstream/ If the test set does not have an auto test capability, test these parameters downstream data rate using a VDSL test set in the VDSL-remote-terminal individually. emulation-mode.

DSL RATES & LEVELS Step 3 Config: VDSL RT, PPPoE Status: Showtime (for 16 seconds) Type: RT/VDSL2 (G.993.2) Run a TDR trace looking for bridged tap. The distance to the open at the end Down Up of the pair should match the open meter distance reading. A typical trace Actual Data Rate: 130000 50000 kbps Attainable Rate: 133591 51907 kbps % Capacity: 97% 96% with an open at 2245 ft is shown in Figure 20. Signal/Noise Ratio: 7.3 6.3 Line Attenuation: 1.8 2.7 dB Signal Attenuation: 0.1 2.7 dB If a bridged tap is identified as in the trace shown in Figure 21, locate and Transmit Power: 11.2 7.3 dBm Fwd Err Corr: 0 0 remove it. They are normally found at a spliced lateral or left-in drop. In the CRC Errors: 0 0 trace below, the first cursor from the left is at the beginning of the bridged RATES & BINS TCP/IP DSL LEVELS GRAPH TESTS CONFIG tap and the next cursor is at the end of the lateral.

Figure 19. Sample VDSL test set screen

The actual data rate may be lower than the attainable rate due to circuit design or company sales strategy. If it won’t sync at the DSLAM, the tech confirms that he or she is on the right circuit and it has been provisioned.

30 Guide to the Last Mile www.megger.com Guide to the Last Mile 31 Layer 2 digital data throughput that give the tech some hints at which Layer TDR <0 ft Gain = 0 2500 ft> <0 ft Gain = 0 893 ft> 1 faults may be causing the data throughput to be reduced. Among them are:

Bit-per-tone Without going too deeply into it, the VDSL signal is broken up into VOP=0.700 2245 ft sub-channels or tones. Depending on the band plan and profiles used VOP=0.670 127.6 264.0 ft Delta Cursor (mentioned earlier), the number of these tones varies. Each of these tones LESS MORE ZOOM SETUP LESS MORE can carry a maximum of fifteen bits. If all is well, they do carry fifteen bits CABLE CABLE CABLE CABLE ZOOM SETUP and the maximum data throughput is achieved. If all is not well, some Figure 20. Figure 21. of the tones may carry no bits while others do carry fifteen bits. At what Step 4 frequencies this happens can help the tech determine the cause. In other Check the noise floor on the spectrum analyzer to 8, 12, 17 or 30 MHz cases, the number of bits carried in most of the tones is reduced. (depending on the VDSL2 band plan used). The noise floor should be≤ 60 dBm across most of the spectrum as in Figure 22. Signal-to-noise ratio

SPECTRUM This is simply the ratio of the power of the usable signal being transmitted 12 MHz dBm to that of an undesired signal (noise). The goal is to have high signal power -20 -30 and low noise power; a high signal to noise ratio. If data throughput is -40 reduced, is the cause low signal power or high noise power or, perhaps, -50 -60 both? The answer can help the tech find the source of the reduced -70 throughput. -80 -90 -20 dBm1.42 MHz VDSL test sets typically display these in what is called a “Bins Graph” as seen LESS MORE VBAND WBAND in Figure 23. The screen shows that each of the tones (also referred to as FREQ FREQ TONE TONE “bins”) is carrying 15 bits and that the signal to noise ratio is about 50. This Figure 22. is not likely unless the tech is right at the DSLAM.

DSL BINS GRAPH

55 Step 5 50 45 Verify that there is no impulse noise. Connect the circuit to the pair. 40 35 SNR 30 25 20 15 Step 6 10 5 Go the customers premise. Connect to the pair with a VDSL test set. Note 10 the actual upstream/downstream rates. A typical screen is shown previously BITS 20 40 60 80 100120 140160 180200 220240 32 0 0.0dB 138kHz in Figure 19. BIN BITS STREAM SNR FREQUENCY RATES & BINS TCP/IP DSL LEVELS GRAPH TESTS CONFIG Troubleshooting VDSL

Slow service/reduced data throughput Figure 23. Sample of “bins graph” We have discussed the faults on the copper pair (Layer 1) that could cause reduced data throughput. VDSL test sets have capabilities of measuring the

32 Guide to the Last Mile www.megger.com Guide to the Last Mile 33 Consider the screen in Figure 24. It shows signal-to-noise ratio falling as Important Factors in Selecting a VDSL Test Set the frequency increases. The tones-per-bin number is less than 15 in many n Fast processor to speed up (reduce/quicken) individual and overall testing places. This is a circuit that is likely too long. (If tones are empty [no bits] the times. tech can determine if it is in the upstream or downstream band by referring to the band plan. Some empty tones will not affect the overall circuit n Lightweight unit for ease of use and transport. performance.) n Rechargeable batteries, ac or dc power.

ADSL BINS GRAPH n Simple menus, intuitive design. 55 50 45 n 40 Comprehensive test suite to reduce the number of required test sets. 35 SNR 30 25 n Complete range of testing for physical layer verifications (POTS) through 20 15 10 VDSL signal transmissions. 5

10 Many of the advanced requirements are: BITS 20 40 60 80 100120 140160 180200 220240 n 32 0 DOWN 0.0dB 138kHz To check and display both AC and DC volts simultaneously for instant BIN BITS STREAM SNR FREQUENCY RATES & BINS TCP/IP ADSL assessment of potentially hazardous and acceptable voltage limits. LEVELS GRAPH TESTS CONFIG n A stress test to reveal pair imbalances or intermittent conditions that disrupt voice or digital signals. Figure 24. Sample of signal-to-noise ratio n Single- and dual-trace TDR with a zero (0) dead zone to locate the various Dips in bits-per-tone indicate interference. Checking the frequency of the conditions indicating impedance mismatches: shorts, opens, splits and interference often makes it possible to indentify its cause. If the bits-per- recombined pairs, wet cable, bridged tap, single-wire splice connectors, tone numbers are reduced and it is determined that noise is not the cause, connecting wire in a cross connect. Dual-trace TDR allows comparison of the cause could be DC faults such as bridged taps, wet sections or other traces from good and faulted pairs to improve pinpoint fault locating. impedance changes. n Resistive Fault Locator to display distance readings of pair or single Very low or non-existent bits-per-tone in the high frequency band indicate conductor faults <2 MΩ. the presence of a long loop. If there is a major dip in bits-per-tone, the most n Spectrum analyzer (voice band and wide band up to 30 MHz) to identify likely troubles are bridged taps, wet sections or other changes in impedance. the frequencies and potential sources of noise that affect services.

If bits-per-tone numbers are low across the entire bandwidth, the cause is n Auto tests, user-defined test sequences, to simplify and standardize basic likely DC faults such as shorts or grounds. testing, and store for reference.

A major dip in bits-per-tone indicates AC trouble on the loop. To verify the n Incremental pair test that uses auto test function to incrementally record type of AC trouble, the tech compares the signal-to-noise to bits-per-tone. each test result with pair ID, date and time stamp; has the ability to store a If they occur at the same frequency, the degraded throughput is most likely large number of tests for later recall and follow-up testing; monitor pair’s due to a transmission influence. condition over time. n VDSL2+ golden modem to verify DSLAM connectivity and internet

connection via multiple data points: actual and attainable data rate, percent capacity and signal to noise ratio up and down stream, S/N ratio and bits-per-bin graph, ping statistics, and more.

34 Guide to the Last Mile www.megger.com Guide to the Last Mile 35 Available from Megger HT1000/2 Copper Wire Analyzer n Noise finder via a 30 MHz spectrum analyzer n 7 user selectable auto tests n Incremental pair test program n 200 pair pre-post test storage n AC or DC power n USB port downloads updates and uploads test results

The HT1000/2 is a high performance, full feature, hand held instrument designed to provide copper wire provisioning and maintenance technicians with the most critical tests at the touch of a button. Durable and water resistant, it is equipped with a highly effective 1/4 VGA The HT1000/2 is the ideal LCD screen and a powerful backlight designed to fully-featured instrument on make testing and troubleshooting easier in all work the market today for physical layer and VDSL testing environments.

The on-screen menu launches most tests with a single keystroke.

Super Stress™ reaches beyond standard longitudinal balance testing, identifying even hard-to-find short loop unbalances.

Dualtrace TDR is standard, with 12 trace storage and intermittent fault location.

The HT1000/2 has user-selectable auto tests with an incremental pair testing process.

Test for DC and AC volts at the same time, no need to switch between separate screens.

Download updates and upload test results quickly and easily via the integrated USB port.

36 Guide to the Last Mile 38 Guide to the Last Mile Megger is your source for testing needs

Megger offers the following high-quality instruments for testing telecommunication systems. n Insulation Resistance Test Equipment n Battery Test Equipment n Ground Resistance Test Equipment n Time Domain Reflectometers n Multimeters/Clampmeters n Low Resistance Ohmmeters

Visit our website for additional information at www.megger.com.

UNITED STATES CANADA 2621 Van Buren Avenue Unit 106-550 Alden Road Norristown, PA 19403 Markham, ON L3R 6A8 866-254-0962 800-287-9688

VDSL_en_V01